Chapter 5: Six causes and four conditions, another algorithm for relations¶

The previous chapter ended with a distinction. Fascicle 157 says that a characteristic and what it characterizes necessarily obtain one and the same fruit; the characteristic follows what it characterizes, is inseparable from it, and neither precedes nor follows it in time. Acquisition and what is acquired do not have the same fruit, do not necessarily follow one another, are not inseparable, and may occur before or after one another. Both tie two dharmas together, but the ties differ in tightness. To explain this difference, possession or persistence alone will not suffice. It takes an entire account of causes and conditions. Sarvāstivāda has one, and the Southern tradition has another. Their shapes differ greatly. This chapter compares them.

The terms need not be listed again. Part I has already set out the definitions of the six causes, four conditions, and five fruits; Essay 14, Chapter 5 has set out the twenty-four conditions. This chapter will not do so a second time.¹ It asks something else. With twenty-four entries on one side and six plus four plus five on the other, where does the numerical difference come from, and where does its cost fall?

First, look at the shape. The Southern tradition's twenty-four entries form a flat list. Each is a type of relation; none is subsumed under another, and one can begin anywhere. Sarvāstivāda's system groups its entries into two levels. The four conditions encompass everything. Of these, the causal condition subsumes five of the six causes; the remaining efficient cause is distributed among the other three conditions. The levels are clear, and the number of entries is reduced by more than half. It looks as though the same business has been put into better order.

Yet this order was fitted together afterward. The four Chinese characters 六因四缘, "six causes and four conditions," occur together just once in the 200 fascicles of the Mahāvibhāṣā, in fascicle 23. And the passage is not a discussion of causes and conditions. It defines an Abhidharma master: one must understand the meaning of the Abhidharma Piṭaka through fourteen matters, of which the six causes and four conditions are only the first two. The same passage records an alternative account that reduces the matters to seven.² The four Chinese characters never occur together in the thirty fascicles of the Kośa or the eighty fascicles of the Nyāyānusāra.^{3 4} All three lists are there individually. Their presentation as a single system of three tiers is a later convenience. Chapter 2 recorded that the table of five categories and seventy-five dharmas does not appear in the Vibhāṣā; Chapter 3, that the four Chinese characters 三世实有, "real existence in the three times," do not appear there; Chapter 4, that "acquisition preceding the dharma" and "acquisition accompanying the dharma" do not appear there. Here is the fourth example.

The treatise gives its own account of their origins. The discussion of the six causes in fascicle 16 opens with a very plain statement: "These six causes, however, are not taught in the sūtras. The sūtras speak only of four conditions. Wishing now to distinguish the conditions by means of causes, we therefore speak of these six causes."² The six causes were devised to subdivide the four conditions in the sūtras. The treatise itself knows this. The other end is interesting: when fascicle 121 discusses the five fruits, it says instead, "The sūtras, however, speak of five kinds of fruit."² Three lists: the two ends claim a scriptural origin; the one in the middle admits to being newly made.

The weight of each list, together with the origin the treatise claims for it:

Having admitted it, the treatise tries to take it back. In the same fascicle 16, someone next says that the six causes too were taught in a sūtra, but that the sūtra has been lost. The Ekottarika Āgama originally increased from one dharma to a hundred; now only one to ten remain. The Venerable Kātyāyanīputra and others, through the power of vow-knowledge, saw the passage in the sūtras on the six causes and composed the treatise on that basis. Another example follows: on the day the arhat Śāṇakavāsin entered parinirvāṇa, 77,000 Jātaka sūtras and 10,000 Abhidharma treatises vanished together. If the passing of one master took so much away, how much more must have gone with the hundreds and thousands of masters who followed him?² Someone else says that although no single sūtra presents all six causes in order, they are scattered through the sūtras. One scriptural passage is then matched to each cause to make up the six.² This book holds that the structure of this defense is itself evidence: when a claim needs something already lost to guarantee it, the person making the claim knows that it is new.

There are also two accounts of the grouping. Do causes subsume conditions, or conditions causes? Fascicle 16 gives two answers. One says that the two subsume one another, depending on the matter at hand. The other says that conditions subsume causes but causes do not subsume conditions, because the immediately preceding condition and the object condition are not among the six causes.² This is no small difference. On the latter account, two of the four conditions have no counterpart at all among the six causes. The two levels thus overlap rather than fitting one inside the other. Another seam is more conspicuous. Fascicle 16 says that the six causes were devised to display four fruits: the associated and coexistent causes display the fruit of effort; the homogeneous and universal causes, the fruit of similar outflow; the maturation cause, the fruit of maturation; and the efficient cause, the predominant fruit. Once understood, this is as plain as an āmalaka fruit in one's palm.² The count here is four fruits, not five. The fruit of disconnection has no place in this table of correspondences. Six, four, and five do not mesh with one another. The seams were smoothed over later.

Figure terminology and omitted wording. The “antecedent” and “objective” conditions are the immediately preceding condition (等无间缘) and the object condition (所缘缘). The “dominant” condition is the predominant condition (增上缘). The “fruit of human effort,” “outflowing fruit,” “dominant fruit,” and “fruit of disjunction” are the fruit of effort (士用果), the fruit of similar outflow (等流果), the predominant fruit (增上果), and the fruit of disconnection (离系果), respectively. “For everyone besides” refers to all the other dharmas. The Chinese box for the fruit of disconnection also says: “The unconditioned has neither cause nor fruit.” The closing description says that this framework uses the vocabulary of real existence in the three times (三世实有), not merely the three times. “Function” means activity (作用), and “Part” refers to this essay.

People already asked at the time whether this apparatus was real. Fascicle 55 begins its discussion of the four conditions by recording the opposing side. Some hold that conditions have no real nature; the Dārṣṭāntikas are among them. Their reason comes from scripture. A sūtra says that ignorance conditions formations, yet ignorance has no distinct characteristic, whereas formations do. How can a dharma without a distinct characteristic condition a dharma with one and itself really exist? The Venerable puts it more plainly: masters designate the names of conditions according to their conceptions; these have no really existing nature.² The point is that the names of conditions are arranged to suit the theorist's ideas. There is no thing in the world called an immediately preceding condition. The Vibhāṣā's reply deserves attention: if the natures of conditions do not really exist, no dharma has a real nature, because the four conditions encompass all dharmas.² This retreats in order to advance, but retreats too far. It amounts to admitting that conditions have no substance apart from dharmas: they are the dharmas themselves, arranged by role. We will return to this at the end of the chapter. The treatise adds that if the natures of conditions did not really exist, there could be no distinction between the higher, middle, and lower grades of awakening. Ghoṣaka goes further: a teacher could not improve a pupil's wisdom from inferior to superior, and a pupil could never become a teacher.² This book holds that neither of the latter two points constitutes an argument. They describe what would happen without causal relations. But the Dārṣṭāntikas are not denying causal relations; they are denying that relations have separate entities of their own.

We should also record whom this apparatus was aimed at. Fascicle 16 says that the six causes were devised to refute theories of no cause and wrong causes: non-Buddhists hold either that dharmas arise without causes or that they have an irregular cause.² At the end of its section on the four conditions, fascicle 7 of the Kośa attacks this line more fiercely. Dharmas do not arise from a single cause such as Īśvara, a self, or primordial nature. If they really came from one cause, everything should arise at once, without a sequence of before and after. It ends by flinging out, "Bah! What use is such an Īśvara?" and, in passing, takes apart the name Rudra.³ The conclusion is that "a single cause producing dharmas is definitely impossible."³ The first job of a theory of causes and conditions is to block the road to a single cause, before explaining how the world works. The dispute over the number of entries belongs in this setting. The Southern tradition and Sarvāstivāda are doing the same thing: spreading the causes of an event out into many entries. They spread them differently.

Fewer entries therefore do not mean simpler relations. The difference between twenty-four and six plus four has not disappeared. It has been pushed elsewhere. What follows compares four points one by one, all explicitly recorded in Essay 14, Chapter 5. Once the four are compared, it will be clear where the difference has gone.

First, the predominant condition has the same name but almost the opposite meaning. Sarvāstivāda's predominant condition means nonobstruction. Every conditioned dharma has all dharmas other than itself as its efficient cause, because the other dharmas did not stand in the way when it arose. The Kośa supplies an analogy: because the ruler of a country does them no harm, all its people say, "Through our ruler we have peace and happiness." It adds that even a powerless king counts.³ Someone presses the point: if all dharmas are efficient causes of one another, why do they not arise at once? The answer is that this is solely through nonobstruction, without a direct power to act upon their arising.³ The Southern tradition's predominant condition (adhipati-paccaya) means whichever of desire, mind, effort, and wisdom takes the leading role, heading an undertaking.¹ One means not standing in the way; the other means taking the lead. The two are almost opposites.

This deserves another sentence. The broadest compartment in Sarvāstivāda's system is defined by a negative, and this is precisely the compartment that swallows up the efficient cause among the six causes. The broadest is the emptiest. As taxonomy, this is a choice, not a defect. Making nonobstruction a compartment amounts to admitting that most relations need not be subdivided: it is enough to ensure that they do not obstruct one another. The Southern tradition spends its effort on subdivision; Sarvāstivāda spends it elsewhere. The next point shows where.

Second, the homogeneous and universal causes versus the Southern tradition's repetition condition. Both concern the successive production of like by like, but the lengths they cover are vastly different. The Southern tradition's repetition condition (āsevana-paccaya) is confined to the seven impulsions in a single cognitive process. The preceding impulsion strengthens the next; once outside that process, the condition has no further business.¹ Sarvāstivāda's production of like by like holds across lives. Wholesome roots in a previous life are homogeneous causes of wholesome roots and associated dharmas in their own realm in a later life; past wholesome roots are homogeneous causes of future and present ones, and present wholesome roots of future ones.² One extends across innumerable lives; the other is confined to seven moments.

Holding across lives is possible only through real existence in the three times, and the treatise says so outright. The discussion of the homogeneous cause in fascicle 17 asks, as usual, why this discussion is undertaken. The answer is: "To refute other schools and display the correct reasoning. Some hold that past and future dharmas have no real substance, or that present dharmas are unconditioned, or that a thing's own kind is its homogeneous cause. To refute these various divergent attachments, we demonstrate the real existence of past and future dharmas."² The first reason for establishing the homogeneous cause is to prove that past and future dharmas really exist. This book need not discover the connection from outside: the treatise itself writes it down as its motive for making the rule. Chapter 3 and this chapter thus treat two ends of one matter, not two separate matters. Real existence in the three times makes causes across lives possible; causes across lives in turn become the most important use of real existence in the three times. This book considers this the most important structural discovery in the essay: the two claims support each other. Remove either, and the other immediately collapses.

The pillars support each other, and the costs pass between them. Since a homogeneous cause requires a before and after, the future becomes a problem. Future dharmas abide in disorder, without a before and after. If there were no before and after, they would have to be causes of one another; if causes of one another, then fruits of one another. Yet being causes and fruits of one another contradicts reason.³ The conclusion is that there are no homogeneous causes in the future.² The cost of this conclusion is considerable. To accommodate it, fascicle 17 works through earlier passages in the Vijñānakāya, Prakaraṇapāda, and Prajñaptipāda one by one. In the end, faced with the Prajñaptipāda's statement that dharmas are fixed in four respects, it can only answer, "It should be stated thus: in the future there are no homogeneous causes." For several other passages, it answers, "This answer was given with an implicit reference to the final position."² An answer given implicitly means that, read literally, those texts do not fit the school's position: they must be assumed to carry another, hidden meaning. To preserve one cause, the treatise rereads three pāda treatises.

The same reason also bars the immediately preceding condition from the future: future dharmas abide in disorder without a before and after, so the future has no immediately preceding condition.³ The next question is how the Buddha knows what will arise in the next moment. The Kośa records three answers: he directly knows it by inference from past and present dharmas; he knows it through vow-knowledge; or there are, within a sentient being's body, portents of future fruits and causes, a particular kind of dissociated formation within the formations aggregate.³ The third answer warrants a pause: to patch one sentence, another formation dissociated from mind is invented. Chapter 4 counted acquisition, nonmanifest material form, group homogeneity, the life faculty, and intermediate existence. Here the same procedure is used once more in the theory of causes and conditions. Vasubandhu accepts none of the three answers. He says directly that the Sautrāntika masters' account is good: the Blessed One knows all dharmas simply by directing his attention, neither by inference nor by divination.³

One further restriction on this condition should be recorded. The immediately preceding condition applies only to mind and mental factors, not to material dharmas, because their quantities are unequal. The Venerable gives examples: burning a great heap of rice straw leaves only a little ash, so much produces little; a tiny seed grows into a banyan tree, so little produces much.³ Mind and mental factors, by contrast, are equal from one moment to the next: one mind, one feeling, one perception, no more and no less. Only they can therefore form an uninterrupted series. Essay 14 said that the Southern tradition's contiguity condition presupposes a shared clock.¹ Sarvāstivāda presupposes the same clock, with the additional restriction that it governs mind alone, not matter. The continuity of material dharmas has another account, which belongs to Chapter 6. The object condition goes the other way, so broad as to be almost unrestricted. The Kośa says that if a dharma has ever been the object of another dharma, there is no time when it is not its object. Similarly, fuel is called what is burned even when it is not burning, since its characteristics are no different.³ Being cognized and being cognizable are made one thing. Here again, real existence in the three times provides the support.

Third, the postnascence condition has no place in this system. The Southern tradition establishes the postnascence condition (pacchājāta-paccaya) to address the mutual support of body and mind from moment to moment: mind arising later supports the continued existence of the material body that arose earlier.¹ Sarvāstivāda's causes all point from earlier to later; there is no such entry. The side of conditions, however, leaves an opening. Fascicle 55 says that future personality view, in relation to past and present personality view, serves as both object condition and predominant condition if it is taken as an object; if not, it serves only as predominant condition.² A later dharma can be the object of an earlier one, because future dharmas really exist, and what really exists can be taken as an object. Sarvāstivāda thus has backward relations, but puts every one of them among conditions and none among causes. And this backward relation is asymmetrical. An earlier dharma is a predominant condition of a later dharma, and the later one is also a predominant condition of the earlier, yet only the later dharma counts as the earlier one's predominant fruit; the reverse does not hold.² The same condition can bear fruit when it runs with time. Against time, only a status in name remains.

Fourth, the four entries provided for dharmas that have ceased. The Southern tradition establishes decisive support, absence, disappearance, and karma at a different moment to answer one question: how can a dharma that has already ceased still support a present dharma?¹ Sarvāstivāda needs only one sentence here: the dharma-substance always exists; what ceases is activity, not substance. Essay 14 judged that the Southern tradition pays in the length of its list, Sarvāstivāda in its ontology. Checking that judgment from Sarvāstivāda's side, we must ask: is ontology really the only place it pays?

The answer is no. Saying that the dharma-substance always exists does not settle the bill in one payment. There are installments. The first is that the future must be written as a disorderly heap, inaccessible to both the homogeneous cause and the immediately preceding condition. The second adds a new formation dissociated from mind to preserve the Buddha's omniscience. The third is the heaviest: every cause and every condition needs a separate time in which it is active. Fascicle 21 classifies the six causes by time. The associated, coexistent, and maturation causes extend through all three times; the homogeneous and universal causes belong only to the past and present; the efficient cause extends through all three times and also beyond time.² The Kośa's verse also fixes the positions of activity: "Two causes act at the moment of ceasing; three causes at the moment of arising." The other two conditions act in the opposite way.³ Projecting and giving a fruit each have their own times too. Five causes project a fruit only in the present. The associated and coexistent causes also give a fruit in the present; the homogeneous and universal causes give a fruit in the past and present; the maturation cause gives a fruit only in the past.³ A timetable accompanies the table of causes and conditions, and is longer than the table itself.

With the four comparisons complete, we can set them side by side:

The cost is most exposed in fascicle 16. There the treatise gives exhaustive definitions of the six causes: the associated cause is all mind and mental factors; the coexistent cause, all conditioned dharmas; the homogeneous cause, all past and present dharmas; the efficient cause, all dharmas.² Fascicle 55 says that the four conditions encompass all dharmas: the causal condition encompasses all conditioned dharmas; the immediately preceding condition, all past and present mind and mental factors except an arhat's final mind and mental factors; the object condition and predominant condition each encompass all dharmas.² Both lists rearrange the same heap of dharmas by role. Neither adds a new thing. Both leave the same two holes: an intrinsic nature does not serve as an efficient cause of itself, nor do conditioned dharmas serve as efficient causes of unconditioned dharmas, or unconditioned dharmas of unconditioned dharmas. The answer for the four conditions is identical.^{2 3} Apart from these two exceptions, no dharma can stand outside causes and conditions. Yet someone immediately points out that this still does not fully explain the matter, because we do not know in what position which dharmas are causes of which.² The two Chinese characters 何位, "in what position," are the hinge of the chapter. Counting all the dharmas does not yet explain causality. One must also say which compartment of time each occupies.

The question "in what position" receives answers in three places: fascicle 21 divides by time, the Kośa's verse by projecting and giving a fruit, and fascicle 16 by the fruit displayed. Superimposed, the three look like this:

Only four kinds of fruit fall into the last column; the fruit of disconnection has no place there. This is precisely why fascicle 16 finishes counting at four.

Now consider fruits. The Southern tradition places fruits within its classification of mind. Resultant consciousness is a class of consciousness distinguished by its activity; its identity is determined by its position in a cognitive process.¹ Sarvāstivāda places fruits within a classification of causal relations. The five fruits stand opposite causes; they are not a kind of mind. The two schools put the same matter in different drawers, and that choice itself has consequences.

Putting fruits in the drawer of relations requires an answer to how many kinds there are, and the number was fought over. Fascicle 121 says that Western masters establish four kinds besides the five fruits: establishment, application, aggregation, and cultivation. The masters of Kashmir say that these four are included in the earlier five, all falling under the fruit of effort and the predominant fruit.² The Kośa records the same four, likewise subsuming them under effort and predominance.³ Add the earlier point that fascicle 16 counts only four fruits, and it becomes clear that five is the result of an adjudication, not a ready-made fact. The side that puts fruits in the drawer of mind need not answer this question. Resultant consciousness is a class of consciousness; how many classes of consciousness there are belongs to another taxonomy.

The fruit of disconnection is the part that will not join up. Cessation through discernment is unconditioned. The unconditioned has neither cause nor fruit, so on what grounds is it called a fruit? The Kośa handles this by distinguishing the path's function with regard to acquisition from its function with regard to cessation. Its function with regard to acquisition is called producing; with regard to cessation, realizing. The path is not a cause of cessation, yet cessation through discernment can be called a fruit of the path.³ The Sautrāntikas immediately object: the unconditioned is not a cause; no sūtra says that a cause is unconditioned.³ The answer is that although no sūtra says so, none forbids it either. Moreover, innumerable sūtras have now vanished. How can one be certain that no sūtra says so?³ The lost sūtras make their second appearance, this time to silence an objection. This book holds that the embarrassment of the fruit of disconnection lies not in the distinction between realizing and producing but in a classification with only one compartment, cause and fruit. Since the relation between cultivating the path and nirvāṇa is not production, it ought to have a separate designation. But there are only five designations, so it is written down as a kind of fruit.

The pairing of the fruit of effort and the predominant fruit is where this classification gets things right. The same accomplished undertaking is both a fruit of effort and a predominant fruit for the person who carries it out, but only a predominant fruit for the person who enjoys it. The Vibhāṣā takes farming as its example: crops are a fruit of effort for the farmer, but only a predominant fruit for the person who eats. The Kośa takes a craftsperson: the same finished object has both names in relation to its maker, but only one in relation to someone else.^{2 3} The name of the fruit changes with the end at which one stands. This is precisely what a taxonomy of relations ought to be able to do. A taxonomy of mind cannot say this much. It can say that a consciousness is resultant consciousness, but cannot say whose doing it results from.

The reverse is equally true. In the Southern tradition, resultant consciousness has a definite place in the cognitive process. To call it resultant consciousness is to say at which step it appears. Sarvāstivāda can say only that a fruit of maturation is unobscured and indeterminate, and must belong to a sentient being. The receptacle world does not count as a fruit of maturation, because it is shared, whereas a fruit of maturation cannot be enjoyed in common.³ Each vocabulary has its blind spots. This book's position is that neither is right or wrong here. The issue is how a taxonomic choice shapes the questions that remain possible. Put fruits in mind, and questions turn toward the sequence of cognitive processes. Put fruits in relations, and they turn toward how many kinds of fruit there are and whether the unconditioned counts as one.

The final internal tension concerns the coexistent cause. The association of mind and mental factors, the four great elements as coexistent causes of one another, and mind together with the bodily and verbal karma that follows mind all require cause and fruit to be simultaneous. The criterion is that whatever are fruits of effort of one another are coexistent causes of one another.³ In passing, a common claim needs correction. Although acquisition and what is acquired often arise together, they do not count as coexistent causes. The Kośa expressly says that they do not invariably occur together: acquisition may arise before, after, or together with what is acquired. Thus, though called coexistent, they do not become causes through the coexistent cause.³ This exclusion is telling. The acquisition of Chapter 4 was made precisely so that it could be separated in time from what is acquired. Acquisition preceding, accompanying, and following the dharma are three names for that separation. The design that frees possession from the present is also what disqualifies it from simultaneous causality. Atoms of resistant derived matter are excluded too, because they do not belong to one and the same fruit.³

The associated cause and the coexistent cause are in fact two aspects of one matter, and the treatise knows it. Asked how they differ, fascicle 16 gives several explanations in succession: being companions is the associated cause, having one fruit is the coexistent cause; holding a staff is the associated cause, accomplishing something after taking hold of it is the coexistent cause; joining hands is the associated cause, crossing a torrential river after joining hands is the coexistent cause; following one another in accord is the associated cause, being inseparable is the coexistent cause.² The same mind and mental factors are associated causes when seen as companions, coexistent causes when seen as accomplishing one undertaking together. This is the chapter's second encounter with one matter going into two drawers, except that this time both drawers are in Sarvāstivāda's own house. One further point: these are precisely the least secure of the six causes. Two scriptural statements are assigned to them as evidence: "with seeing as root, faith, realization, and knowledge are associated," and "conditioned by eye and visible form, eye-consciousness arises; the aggregation of the three is contact, and feeling, perception, and volition arise together." Both concern the simultaneous occurrence of mind and mental factors. Yet whether mind and mental factors occur simultaneously is precisely what Sarvāstivāda needs to prove. Using it as evidence is circular. Modern scholarship has specifically examined the difference in level between the lists of six causes and four conditions, the dispute between the Kashmir and Western masters, and the controversy over the coexistent cause.⁵

Simultaneous causality is needed because dharmas are too weak. The reason given in fascicle 16's section on the associated cause is that conditioned dharmas are weak by nature and can accomplish their work only by depending on one another. It supplies three analogies: two bundles of reeds standing by leaning on each other, many ropes combined to pull a large log, and many people joining hands to cross a great river.² Fascicle 55 expands the point in its discussion of the four conditions. Conditioned dharmas are weak in their intrinsic nature, lack autonomy, depend on others, have no activity of their own, and cannot follow their own wishes. They therefore arise from four, three, or two conditions. Not even one condition can produce them alone, much less no condition. They are like a sick and weakened person who needs four, three, or two people supporting him to rise and stand.²

There is a deep tension here that deserves separate attention. Every dharma really exists and its intrinsic nature always exists: these were the conclusions of Chapters 1 and 3. Yet the same dharmas, as conditioned dharmas, are weak, have no activity of their own, and cannot follow their own wishes. Real existence and powerlessness stand together. This book holds that this is the precise weight of "intrinsic nature" in Sarvāstivāda. It guarantees only that a dharma has its own determinate characteristic, never that it can accomplish anything alone. Essay 14 recorded the Southern tradition's opposite move: conditional power is intrinsic to the conditioning dhamma, like a chili's heat; the relation is taken into one of its terms.¹ Sarvāstivāda cannot take that route, because its dharmas lack power from the outset. Power can come only from arrangement. Yet the arrangement is itself no new thing: the six causes and four conditions merely rearrange the same heap of dharmas. If power is neither in dharmas nor outside them, only one place remains to put it: the compartment of time in which a dharma stands.

This is where Vasubandhu strikes. He says that worldly causes and fruits, such as seed and sprout, have never been seen to occur simultaneously. Someone offers lamp-flame and lamplight, sprout and shadow, as things that are simultaneous and causes and fruits of one another. He will not accept this. One must first distinguish whether flame and light are causes of one another, or whether the coming together of prior causes and conditions makes both arise at once.³ He then invokes the logicians' criterion: if that exists when this exists, and that does not exist when this does not, then this is certainly the cause and that certainly the fruit.³ By this criterion, simultaneity need not rule out causality. But being causes and fruits of one another makes no sense. His exact words are: "The principle of causes and fruits arising together may provisionally be granted; how can they be said to be causes and fruits of one another?"³ Sarvāstivāda offers the analogy of three staffs standing by leaning on one another. He dissects it: do the staffs support one another because they have been set upright together, or have other prior causes and conditions made them stand together? Besides, ropes, hooks, and the ground hold them together too.³ The Sarvāstivāda side answers that homogeneous causes and the like are also present there, so the coexistent cause is established.³

That answer concedes a step. Rescuing a simultaneous cause with a prior cause admits that simultaneity cannot support itself alone. This book holds that the real difficulty is mutuality, not simultaneity. Simultaneous causality in one direction is nothing unusual; it is being causes of one another that is awkward. Sarvāstivāda must insist on mutuality because it has already made the characteristic and what it characterizes, mind and mental factors, into separate entities. That work was done in Chapters 2 and 4. Once they are independent things, a rope must bind them. And if this rope has a before and after, a gap opens where it binds. The system cannot bear such a gap. The coexistent cause patches a problem that pluralistic realism has made for itself.

Saṅghabhadra later reworks this answer. Fascicle 15 of the Nyāyānusāra answers the same objections, but changes the point of attack. Instead of appealing to the presence of homogeneous causes, he works on two fronts. First, the criterion: he finds "fruits of effort of one another" too restrictive, saying that the criterion of a cause covers too little, because causality between a dharma and its secondary characteristics runs in only one direction. He therefore broadens the criterion: having one fruit can qualify things as coexistent causes.⁴ Second, the reference of the two Chinese characters 互为, "one another": the mutuality he accepts does not mean each individual paired with each other individual. It means that all members of an aggregate, considered in relation to one another, are all causes and all fruits. To ask which is definitely the cause and which definitely the fruit is, in his view, to ask in the wrong place.⁴ He supplies another test: cover the lamp with an object; the light ceases, but the lamp remains. They therefore do not both come from a single cause.⁴ This test is absent from the Kośa passage. There Vasubandhu merely juxtaposes the two possibilities without giving a way to distinguish them. Yet the concession remains. Saṅghabhadra himself writes, "I do not admit that a coarising cause alone can produce dharmas without relying on other causes." The simultaneous level still needs other causes to make it hold.⁴

To close in one sentence: the Southern tradition distributes the burden of relations across entries; Sarvāstivāda distributes it across time. This is no figure of speech. In discussing the four conditions, fascicle 55 assigns each an analogy: the causal condition is like a seed, the immediately preceding condition like clearing a path, the object condition like leaning on a staff, the predominant condition like not blocking the way. It then says that because later-arising dharmas are supported by earlier-arising dharmas as four conditions, they can move through the times, project a fruit, perform karma, and cognize an object.² The first thing gained through this support by causes and conditions is the ability to move through the times, to walk through the three times. A dharma first needs causes and conditions so that it can move; arising comes afterward. This sentence draws together all the chapter's threads. Sarvāstivāda's theory of causality is written in the vocabulary of real existence in the three times. Cancel that premise, and not one term of that vocabulary can be used.

Over to Chapter 6. The discussion above repeatedly used the standard example of the four great elements as coexistent causes of one another. It also mentioned that atoms of resistant derived matter cannot be coexistent causes of one another because they do not belong to one fruit. Both statements presuppose an account of matter: what an atom is, how several atoms count as being in one place, and whether what aggregates is the same thing as the atoms gathered together. Sarvāstivāda gives this account in detail. Here, for the last time, it draws the boundary between nominal and real existence at a place that later proves unable to hold. The next chapter takes up the material end.

References¶